May 06, 2020

Name
Yang Liu

Hometown
Taiyuan, China

Major
PhD, Computer Science           

Faculty Advisor
Yi Shang 

Where did you complete your Undergraduate Study (and additional Graduate Study if Applicable)?
MS, Electrical and Computer Engineering, University of Missouri; BS, Electrical Engineering and Automation, Shanxi University, Shanxi, China

Why did you choose Mizzou for your graduate program?
Mizzou is a great university for me. Mizzou is ranked as a top university and focuses on leading research. In computer science, there are many experienced professors who are performing cutting-edge research, which really aligns with my interest to conduct leading research. I am interested in artificial intelligence, and my advisor Dr. Yi Shang is an expert in this field. Moreover, the Mizzou campus is large and beautiful. Mizzou has the dominate Division I athlete program in the nation. The newly built Rec Center is ranked top Rec Center in the U.S. Also, Columbia has a small-town feel, and it’s a wonderful place to study and focus on research. 

Tell us about your graduate research.  What are some potential outcomes or applications?
My graduate research focuses on artificial intelligence and machine learning. Specifically, I am working on the problem of object detection for small objects. It is not an easy problem, due to the lack of context information and complicated background. In my work, we identified the challenges and solutions for small object detection and adapted current state-of-the-art algorithms on our dataset and created our own algorithms. Given an image, we can generate the location of the objects and the count of the objects. Our work is an application for the Missouri Department of Conservation (MDC). Originally, they took a photo and counted birds manually. With our framework, it can detect and count birds with higher accuracy and requires much less time.

Moreover, instead of detecting objects individually, we proposed a new framework which considers relationships between objects. Our graph proposal network (GPN) combines both semantic information and spatial information. Our results show that compared with the state-of-the-art algorithms, our proposed network can dramatically reduce the false detection for small objects.

What are some accomplishments you’ve achieved at Mizzou that you’re especially proud of?
During my PhD, I have published three papers, one paper is under review and one paper is going to be submitted. My publications are as follows: 

(1) Yang Liu., et al. Performance Comparison of Deep Learning Techniques for Recognizing Birds in Aerial Images. IEEE International Conference on Data Science in Cyberspace (ICDSC) 2018.

(2) Guang Chen, Yang Liu., et al. Digital Image Vegetation Analysis with Machine Learning. International Conference onRobotics and Artificial Intelligence (ICRAI) 2017.

(3) Joseph P. Bernstein, Brandon J. Mendez, Peng Sun, Yang Liu, Yi Shang. Using Deep Learning for Alcohol Consumption Recognition. IEEE Consumer Communications & Networking Conference (CCNC) 2017.

Moreover, I have received several awards during my PhD. In 2016, I become the member of the engineering honor society Upsilon Pi Epsilon. In 2017, I received the outstanding member of Chinese Students and Scholars Association. In 2018, I become the student teaching mentor of Big Data Analytics Summer Experience.

What do you hope to do after you complete your degree?
I hope to join the industry and apply my skills and knowledge to build more exciting products which can improve people’s everyday lives. More specifically, I want to use my AI and machine learning skills and conduct research in various fields, such as e-commerce and self-driving cars.